Arm Unveils New Mobile Computing System in China for AI Era

Chip designer Arm launched its CSS for Mobile 2 compute subsystem at its annual Arm Everywhere China event in Shanghai. The new platform is engineered to meet the demands of the emerging intelligent agent era, integrating CPU and GPU components to boost the performance of AI applications and native graphics on mobile devices.

The CSS for Mobile 2 includes the Arm C2 CPU cluster, Mali G2-Ultra NX GPU, and SI L2 system interconnect. Arm states the platform is optimized for task execution speed, efficient data transfer, and system coordination. The company emphasizes that this system enables device manufacturers to develop customized chips, allowing them to focus on their core competencies and innovation rather than expending resources on foundational technology integration.

The new C2 CPU cluster features C2-Ultra and C2-Pro processors, along with dual SME2 units. This configuration significantly enhances AI model performance and reduces power consumption. According to Arm, latency is reduced and memory access speed is improved, which are critical for the rapid responsiveness of AI agents. The company views the CPU's role evolving from pure computation to a more coordinating and orchestrating function.

The Mali G2-Ultra NX GPU marks Arm's first AI-native GPU. It embeds neural network accelerator cores directly into the GPU's shader cores. This allows for parallel execution of neural network tasks alongside graphics and compute workloads, leveraging a shared memory system and reducing data movement. The GPU supports technologies such as AI-powered image upscaling and enhanced frame rates.

Arm positions these advancements as crucial for the proliferation of intelligent agents in mobile devices. The new compute system aims to address the escalating AI requirements and deliver improved performance and efficiency for future mobile experiences.